Crime Category Prediction Using Advanced Machine Learning and Feature Engineering

Crime prediction plays a crucial role in enhancing public safety and optimizing law enforcement strategies. This study utilizes a comprehensive dataset containing detailed information on criminal incidents, including location, time, victim demographics, and crime characteristics. By leveraging machine learning techniques, we aim to develop predictive models that accurately forecast crime categories. The data preprocessing pipeline includes exploratory data analysis (EDA), feature engineering, and model training. Various machine learning models such as Logistic Regression, LGBM Classifier, and Gradient Boosting Classifier were evaluated for their performance. The results indicate that LGBM Classifier achieves the highest accuracy (95.50%), outperforming other models. Feature selection using SelectKBest marginally improves the performance of Logistic Regression and Gradient Boosting Classifier, while TruncatedSVD reduces effectiveness. Through hyperparameter tuning and model assessment using confusion matrices and classification reports, our findings reveal critical insights into crime patterns. This approach can aid law enforcement agencies in proactive decision-making, resource allocation, and crime prevention.